Patients and tissue specimens

Tumor samples from 86 patients with primary STS, who underwent surgery between 2007 and 2014, were obtained from the tissue bank of our institution and snap-frozen in liquid nitrogen immediately after surgical resection until RNA extraction. Paraffin-embedded specimens were used for IHC staining. To be included in our study, cases should be histologically diagnosed as grade 2 or 3 according to the FNCLCC grading system, with no preoperative chemotherapy. Metastasis, overall (OS), 4-year, metastasis-free (MFS), as well as disease-specific (DSS) survivals, were monitored with a mean follow-up of 45 months (range 23–83 months). Sample acquisition was approved by the Ethics Committee of the Hospital. Written informed consent was obtained from all patients.

RNA extraction and quantitative RT-PCR

Quantitative RT-PCR was used to detect the expression levels of AZGP1 mRNA in 81 cases of tumor tissues, since 5 cases were excluded due to RNA extraction failure. Total RNA was extracted from frozen tissues containing > 80% STS cells by RNA Extraction Kit (QIAgen) according to the manufacturer’s instructions. Quantity and quality of RNA was confirmed by a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). RNA purity was determined by an OD260/280 value between 1.8 and 2.0. For mRNA expression, cDNA was obtained from 2 μg total RNA using Moloney murine leukemia virus reverse transcriptase (M-MLV RT) (Invitrogen, Carlsbad, CA) with oligodT₁₅ primers. GAPDH mRNA was used as an endogenous control to normalize for AZGP1 mRNA expression. qRT-PCR was performed using SYBR® Green PCR Master Mix (TOYOBO) on the ABI 7500 Fast (Applied Biosystems). Data were calculated as relative quantification to GAPDH, based on calculations of 2^−△Ct where −△Ct = Ct (Target) – Ct (Reference). Fold change was presented by the 2^−△△Ct method [13]. Sequences of all primers are listed on Additional file 1: Table S1.

Immunohistochemistry

Immunohistochemical staining for AZGP1 was performed in soft tissue sarcoma tissue microarray (TMA) by a standard two-step method. Briefly, the TMA sections were dried overnight at 37 °C, deparaffinized in xylene and rehydrated through a series of graded alcohol. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 20 min. The slides were boiled in 10 mM sodium citrate buffer pH 6.0 by a pressure cooker for 10 min. After washing three times with phosphate buffered saline (PBS; 0.01 mol/L; pH = 7.4), the slides were incubated with 5% non-fat milk in PBS for 30 min to reduce nonspecific antibody binding. Subsequently, slides were incubated overnight at 4 °C with the rabbit polyclonal antibody against human AZGP1 (Abcam; Cambridge, UK; 1:100 dilution). After rinsing, the slides were incubated with goat anti-rabbit antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) at a 1:100 dilution in PBS for 1 h at room temperature, and stained with 3,3-diaminobenzidine tetrahydrochloride (DAB). Finally, they were counterstained with Mayer’s hematoxylin, dehydrated in graded alcohols followed by xylene. Known immunostaining-positive specimens were used as positive controls and slides immunoreacted with PBS were used as the negative controls.

Western blotting

Total protein from cells was extracted in RIPA lysis buffer and quantified using BCA assay. 20 μg protein from each sample was separated by 10% SDS polyacrylamide gel electrophoresis and electroblotted onto polyvinylidene difluoride (PVDF) membranes (Millipore, Bedford, MA). The membranes were blocked in 5% non-fat milk for 1 h, washed three times with Tris-buffered saline containing 1% Tween 20 (TBST) at room temperature and then incubated overnight at 4 °C with the rabbit polyclonal antibody against human AZGP1 (Abcam; Cambridge, UK; 1:2000 dilution). After washing with TBST, membranes were incubated with secondary antibodies at room temperature for 1 h (goat anti-rabbit IgG, 1:10,000 dilution, Jackson ImmunoResearch Laboratories). Following washing with TBST, immunoreactivity was visualized by enhanced chemiluminescence reagents (Millipore). GAPDH served as internal reference.

Cell culture

Three human STS cell lines (RD ATCC^® Number: HTB-166™, SW982 ATCC^® Number: HTB-93™ and HT1080 ATCC® Number: CCL-121™) were purchased from American Type Culture Collection. RD and HT1080 cells were cultured in RPMI-1640 medium (Gibco, CA, USA) containing 10% fetal bovine serum (FBS; Gibco) at 37 °C with a humidified 5% CO2 atmosphere. SW982 cells were cultured in Leibovitz’s L-15 medium (Gibco, CA, USA) containing 10% FBS.

Vector construction

All constructs were made by standard DNA recombination techniques. The human AZGP1 (NM_001185.3) sequences were amplified by PCR from cDNA using primers listed in Additional file 1: Table S1, and subsequently cloned into lentiviral shuttle vector plenti6 (Invitrogen). For AZGP1 knockdown constructs, two short hairpin RNA (shRNA) sequences, including shRNA150 (target sequence: 5’-GGCTCACTCAATGACCTCCAG-3′), shRNA368 (target sequence: 5’-GTGAGATCGAGAATAACAGAA-3′) and scramble control sequence of 5’-GCTTCGCGCCGTAGTCTTA-3′ were designed and cloned into lentiviral shuttle plenti6-U6 vector.

Cell transfection

Lentiviral constructs were transfected into human HEK 293 T cells (ATCC^® Number: CRL-11268™) with the ViraPower Packaging Mix (Invitrogen) to generate lentivirus. For infection, RD cells were seeded into 6-well plates at a density of 5 × 10⁴ cells/well, and infected with AZGP1 over-expression lentivirus or empty lentivirus as control. HT1080 cells were infected with shRNA lentivirus or scramble lentivirus as control. Antibiotic-resistance cells were selected by 5 μg ml⁻¹ blasticidin (Invitrogen) and used for subsequent experiment.

Wound healing assay

Cell spreading was analyzed using the wound healing assay. RD cell layers at 90% density in 24-well plates were scratched with a sterile 200 μL pipette tip and then washed with PBS. After 48 h, spreading cells were observed under the microphotography. Assays were repeated three times for each clone.

Transwell migration and invasion assay

Cell migration or invasion assay was performed in a 24-well Boyden chamber with or without Matrigel as described elsewhere [14]. The cells on the lower surface of the membrane were stained with crystal violet after fixation with 2% methanol for 5 min. Photographs of four randomly selected fields were taken to indicate cells that migrated to the other side of the membrane, and cell numbers were counted under a microscope at 200× magnification. Each test was performed in triplicate.

Statistical analysis

We calculated OS, 4-year, MFS and DSS using Kaplan-Meier analysis. We defined OS as the period between the date of the definitive surgery and the date of death, the 4-year survival as the period between the date of the definitive surgery and 4 years after, MFS as the interval between the date of the definitive surgery and the appearance of metastasis, and DSS as the date between the date of the definitive surgery and the time of death resulting from the disease itself. The effect of AZGP1 expression on Kaplan Meier survival curves was evaluated by the Log Rank test.

Mann Whitney and Kruskal Wallis tests were used to detect any association between AZGP1 mRNA expression and various pathological features (gender, age, TNM classification, recurrence, metastasis, and 4-year survival) between 2 or more groups, respectively.

Univariate analysis between pathological features (age, gender, AZGP1 expression, tumor size and histological grade) and metastasis or disease-specific survival was determined using Pearson’s correlation analysis. Multivariate analysis between the same variables was evaluated by the Cox regression model. Unpaired student’s T-test was performed to evaluate cell migration/invasion after gene modulation. All analyses were performed with SPSS^® software 23.0 program for Windows^® (SPSS Inc., Chicago, IL, USA). The statistical significance between groups was set at a p-value < 0.05.

Patients with low AZGP1 expression had more metastasis and less 4-year survival

Patients with low expression of AZGP1 had shorter overall survival

According to the median value of AZGP1 expression (0.2014) in STS samples, patients were divided into low and high expression groups. Kaplan-Meier survival analysis showed that patients with low AZGP1 expression had significantly shorter overall survival (OS) than those with high expression (Fig. 1d, 75th percentile was 16 vs. 30 months, p < 0.05).

AZGP1 expression was related to metastasis and disease-specific mortality

To further analyze the relationship between AZGP1 level and metastasis or survival of STS patients, we performed IHC staining analysis in TMA with 86 cases of STS tissue samples (Fig. 2a). Patients were divided into low (negative and weak expression) and high expression groups (median and strong expression). Pearson’s correlation analysis showed that AZGP1 expression negatively correlates with STS metastasis and disease-specific mortality (Fig. 2b; r = −0.218, p = 0.044; r = −0.148, p = 0.034, respectively). Consistent with our findings, univariate analysis showed a higher average hazard ratio (HR) for metastasis and disease-specific mortality (Table 2; HR = 3.731, p < 0.05; HR = 2.481, p < 0.05, respectively) in patients with low AZGP1 expression. The results suggest that patients with low expression of AZGP1 are more prone to metastasis and disease-specific death.

Variables	Metastasis		P Value	Death		P Value
	HR	95%CI		HR	95%CI
Age	1.339	0.545–3.289	0.524	0.725	0.289–1.821	0.494
(≥ 60 yr. vs. < 60 yr)
Gender	0.711	0.287–1.762	0.461	0.902	0.362–2.248	0.825
(Male vs. Female)
Tumor size	2.000	0.786–5.088	0.146	3.286	1.202–8.982	0.020
(> 5 cm vs. ≤ 5 cm)
Histological grade	3.750	1.493–9.420	0.005	1.086	1.362–8.712	0.009
(G3 vs. G2)
AZGP1 expression	3.731	1.770–10.204	0.035	2.481	1.022–6.024	0.044
(low vs. high)

Kaplan-Meier survival analysis suggested that patients with low AZGP1 expression exhibited significantly shorter OS and MFS than those with high expression (Fig. 2c and d, 75th percentile was 15 vs. 30 months for OS, p = 0.046; 6 vs. 18 months for MFS, p = 0.038). Multivariate survival analysis using Cox’s regression model, however, failed to identify AZGP1 expression as an independent prognostic factor (data not shown). Consistent with our qRT-PCR results, these data also suggested that low expression of AZGP1 protein were correlated with metastasis and short survival.

AZGP1 expression in STS cells

We then analyzed AZGP1 expression in three STS cell lines (RD, SW982 and HT1080) by qRT-PCR and Western blot. The results showed that AZGP1 mRNA and protein levels in RD cells were lower than those in HT1080 and SW982 cells (Fig. 3a and b).

AZGP1 inhibited cell spreading, migration and invasion in RD cells

In order to document the effects of AZGP1 on cell movement, we tested the cellular spreading capability by the wound healing assay, and the cell migration and invasion ability by Transwell assay after ectopic expression of AZGP1 in RD cells. As shown in Fig. 3c and d, the expression of AZGP1 was up-regulated significantly after RD cells infection with AZGP1 lentivirus. Following the increase in AZGP1 levels, RD cell spreading decreased compared to that of control cells (Fig. 3e). The migration and invasion of cells over-expressing AZGP1 were also decreased by 62% and 81% respectively, compared with control cells (Fig. 3f–h). These results suggested that AZGP1 over-expression had an inhibitory effect on cell spreading, migration and invasion in RD cells.

AZGP1 inhibition promoted migration and invasion in HT1080 cells

We inhibited the expression of AZGP1 using small hairpin RNA (shRNA) in HT1080 cells. As shown in Fig. 4a and b, the expression of AZGP1 mRNA and protein was decreased by 55% for sh150 and 80% for sh368 compared with the control (scramble oligo). As demonstrated by Transwell assay (Fig. 4c), the number of migrated cells was increased by 3.1 fold (Fig. 4d), and the number of invasive cells was enhanced by 5.2 times (Fig. 4e) after knockdown of AZGP1 expression in HT1080 cells by sh368 lentivirus. These findings were in accordance with those inaugurated from the RD cells experiments, and suggested that AZGP1 inhibition promoted cell migration and invasion.